Shift Happens

A tale of ambition and overreach.


“At the age of six I wanted to be a cook. At seven I wanted to be Napoleon. And my ambition has been growing steadily ever since.”

Salvador Dalì

André Citroën, the French industrial giant, may not have possessed a level of ambition quite as extreme as that claimed by the controversial Spanish artist, but few amongst his peers in the automobile industry could match his boundless energy and determination to lead the way, often eschewing received wisdom and conventional thinking in the process. These attributes brought him fame and fortune, but would also eventually prove to be his undoing.

A salient example of the double-edged sword of Citroën’s ambition and overreach was the Traction Avant of 1934. It was a revolutionary, highly modern and accomplished design in almost every possible way. Citroën’s original plan was for the Traction to be equipped with a newly designed fully automatic transmission, the brainchild of a prolific Brazilian inventor. That plan had to be abandoned so, notwithstanding its long list of innovations in both engineering and construction, the Traction would be saddled with an archaic three-speed manual gearbox for the duration of its long career. Not only was its three-speed configuration(1) increasingly perceived as antiquated as time moved on, it was also baulky in operation, noisy and, in early Tractions, prone to failure as well.

The reason for this unhappy compromise was that the manual gearbox had to be developed in a rushed last-minute programme and made to fit in the housing designed for an automatic transmission that never passed the prototype stage and was finally abandoned only weeks before the Traction’s introduction.

Of all the visible placards above the stands at the Paris Salon only Mercedes-Benz, Rolls-Royce and Lancia survive. And that last one, only just. Images: and

Dimitri Sensaud de Lavaud was born in Valladolid, Spain in 1882, but became a naturalised Brazilian in his youth. His father was French and his mother Russian, which explains his unusual combination of given and family names. Dimitri’s father, Baron Evariste Sensaud De Lavaud, operated a ceramic pipe factory in Osasco, Brazil, to where his family had emigrated from Spain.

As a teenager, Dimitri demonstrated a precocious talent for engineering when he invented a machine using a centrifugal casting process that he claimed would greatly improve the production efficiency of his father’s factory. The elder Sensaud de Lavaud was not immediately convinced, but agreed to try it out and pay his son royalties if it worked. It did, and when Dimitri’s machines were sold to other factories, it was not long before the son became wealthier than his father.

In the early twentieth century, Dimitri moved to the USA where he marketed his centrifual casting process with great success. By the time he moved again in 1920, this time to France, he was a very rich man. Having amassed a fortune large enough never to have to work or worry again, Dimitri indulged his creative passions by inventing.

This slideshow requires JavaScript.

Dimitri was fascinated by both aviation(2) and the automobile, two fields of engineering that were still in their early stages of development, so offered great potential for an inventor. At the 1927 Paris Salon de L’Automobile, Dimitri presented a car of his own design. Apart from the absence of running boards, the outward appearance of the Labourdette-bodied vehicle was, at first acquaintance, pleasingly proportioned but quite conventional.

A separately displayed bare chassis proved that one should never judge a book by its cover, however. Made of Alpax, an aluminium-silicon alloy, the naked chassis allowed the mechanically inclined to admire the hydraulic brakes and independent suspension on all four wheels. According to its constructor, the ‘springless suspension’ by rubber cushions was completely maintenance-free.

Up front was a 5.5-litre inline six-cylinder engine made by the American firm, Continental. It was a proven powerplant, but nothing to get excited about. The car’s real innovation was the manner in which drive was transmitted to the rear wheels via a continuously variable fully automatic transmission. This was a complex and intricate piece of engineering, providing automatic drive through a system of oscillating plateaus(3).

This transmission was quite different from the one that might have been used in the Traction, which was still some years away. Offered for just three years, the expensive Sensaud De Lavaud car found few buyers, but those who knew the man confided that, for him, that never seemed to be the point: as soon as he had realised an idea, he did not appear to be very motivated to monetize it, but rather moved on to new technological challenges(4). In 1925, Sensaud de Lavaud was awarded the Knight’s Cross of the Legion Of Honour by the Academy of Sciences in Paris in recognition of the value of his engineering research.

Around 1930, Sensaud de Lavaud first got in touch with Citroën and, two years later, the manufacturer’s Bureau d’Etudes equipped two Citroëns, a C4 and a C6, with his hydraulic automatic transmission for testing purposes. This new transmission was named ‘Turbine’ by the Brazilian inventor who stated in a surviving technical memo: “The engine directly drives two wheels with 318 blades stirring oil in a sealed crankcase. This oil hits fins connected to the transmission shaft, and by adjusting the speed of the engine by ratio to the slope or resistance of the road we obtain, before reaching the direct grip, an infinite range of speeds, of absolute flexibility”.

During one of the test runs, André Citroën himself took the wheel and was impressed: not only were progress and acceleration extremely smooth, the transmission ironed out the last remaining vibrations transmitted by the ‘Moteur Flottant’(5) rubber engine mounts. It is not difficult to imagine Citroën seeing it as the ideal transmission for the revolutionary front-wheel-drive car he was planning to introduce.

The Bureau d’Etudes noted that there were some negatives, however. Acceleration, while smooth, was quite leisurely, fuel consumption was noticeably higher than with a manual gearbox, production costs were substantial and, most seriously, the Turbine had a worrying tendency to overheat its transmission oil on sustained inclines.

The ‘Turbine’ automatic transmission. Image:

It has at times been suggested that André Citroën, blinded by the potential and novelty of Sensaud de Lavaud’s Turbine, decided right there and then that it would be fitted to all Tractions by 1934. The exclusive contract with Sensaud de Lavaud was, however, only signed on 17th november 1933, well over a year after Citroën’s test drive.

A lot of further fettling and testing took place, but the overheating problem persisted. André Citroën reacted in characteristic fashion to doubts and criticisms concerning the problems with the newfangled transmission raised by none other than André Lefebvre: “You are the engineer, you solve them!”. In light of this, it is interesting to remember that Lefebvre had worked at Voisin before coming to Citroën (via a short and unhappy stint at Renault). Some Voisin models had been equipped with the earlier automatic transmission design by Sensaud de Lavaud(6) and similar problems had arisen that even the talented and resourceful Lefebvre was unable to solve.

At the conclusion of yet another failed test, Lefebvre quipped to a fellow engineer: “Do you like fries? Just put some potatoes inside the transmission- they will be nice and crispy by the time you reach the top of the hill at Meudon”. For many others at the Bureau d’Etudes, the fragile and temperamental automatic transmission only increased the complexity of what was already a car full of new ideas that was being developed in record time. Doubts and objections were likely discussed within the Bureau d’Etudes, but all were, of course, aware whose name was above the entrance to the Quai de Javel plant.

From January 1934 onwards, Traction Avants were produced without transmissions or driveshafts in anticipation of receiving those later. Soon, their storage created logistical problems within the factory.

A month later, during a large meeting presided over by André Citroën on the progress of the Traction, Charles Brull, one the company’s most highly respected engineers, threatened to resign if the Turbine was not fixed or abandoned. André Citroën finally gave in (something not exactly typical of the man) and postponed the release date of the Traction to give his engineers one more shot at getting Sensaud de Lavaud’s transmission production-ready. It was all to no avail and, after a final test with Pierre Michelin and Lefebvre where only one of the five cars in the convoy was able to make it back to the factory, it was clear that the Turbine could simply not be used in the forthcoming Traction.

This slideshow requires JavaScript.

(In the photos above, notice the much smaller shift lever of the Turbine (left and bottom left); in the second photo the engine bay on top is a rare image of a Turbine-equipped Traction pre-production car as it does not have the long connecting rods to the gearbox as seen in the bottom photograph.)

André Citroën finally had to face the facts, and instructed the Bureau d’Etudes to design and build a conventional manual gearbox for the Traction. Not unexpectedly, his engineering team had seen this coming and had already commenced development of just that in secret. Camusat, the Bureau d’Etudes’ transmission wizard, designed a thoroughly conventional three-speed gearbox in a very short time. This was all the more impressive as he was at the time recovering from an accident and his arms were in bandages and plaster.

Lefebvre himself designed the unusual gear linkage system, with long rods connecting the gearshift lever to the gearbox which sat in front of the engine. Its strange appearance soon earned it the nickname ‘Eiffel tower’. Just two weeks after the Turbine was axed, the first Tractions with the manual three-speed box were already undergoing test trials.

Interestingly, testing and development of the Sensaud de Lavaud transmission continued until mid-november 1934, just weeks before André Citroën was forced to file for bankruptcy and hand over the reins of his company to Michelin. It appears that there was still the belief, but possibly held only by André Citroën and Dimitri Sensaud de Lavaud, that the Turbine could be made to work reliably and introduced at a later date. It would not be until 1948, however, that Buick would succesfully introduce its Dynaflow transmission, which functioned along similar lines .

Another no-show in terms of public availability: the V8 engined ’22’. Image:

On 18th April 1934, the Traction was oficially presented to the press after a staggeringly short development time of just over thirteen  months. A large part of the credit for this achievement must go to Lefebvre, who was an extraordinarily gifted engineer as well as quick thinker, but who could also motivate those around him. It also helped that Lefebvre was spared by André Citroën personally from time-consuming and distracting administrative duties.

The whole Bureau d’Etudes staff worked extremely hard, many even giving up their usual summer holidays and often working seven days a week.  The Traction Avant itself was so promising and revolutionary that all who worked on it believed in its concept and went the extra mile. By the same token, the fact that most were acutely aware that the future of the company depended on the success of this car motivated them to do everything they could to make it succeed.

Sadly, André Citroën never witnessed his Traction coming good, dying from stomach cancer in July 1935. The first Tractions had their share of problems, especially with the gearbox and homokinetic driveshafts. Within a short time, however, most(7) issues were fixed and the car became a much loved presence on the streets and autoroutes in France and many other countries. The Traction Avant enjoyed a 23-year long career, changing very little during its lifetime, hence proving the essential ‘rightness’ of its concept.


(1) Despite incessant requests from customers for a fourth gear to be added, Citroën never did, prompting enterprising outside suppliers to offer four-speed gearboxes of their own design for the Traction.

(2) On 7th January 1910, Dimitri Sensaud de Lavaud was the pilot of the first controlled powered flight in Latin America. He had personally designed and built the airplane, including its engine.

(3) A technical explanation of ‘oscillating plateaus’ would be much appreciated, if any of our readers could oblige.

(4) Over his lifetime, Sensaud de Lavaud would be granted patents for around 250 inventions.

(5) A Chrysler invention produced under license by Citroën.

(6) Sensaud de Lavaud’s first automatic transmission had been fitted to some Voisins in 1925 but, as that was very much an upmarket car, its technical delicacies were not widely known to the public. It would also be applied in the prestigious but ill-fated Bucciali in 1928.

Sensaud de Lavaud was not the first to come up with the basic idea: The first torque converter was patented as a propeller shaft reducer in 1904 and built in 1908 by German engineer, Hermann Föttinger. In 1914, the German battlecruiser ‘Wiesbaden’ was fitted with two hydraulic converters of his design.

(7) As mentioned earlier, the Traction’s manual gearbox would continue to be the car’s Achilles’ heel. The reliability problems were resolved, but it remained a crude device, unworthy of the Traction Avant.

Author: brrrruno

Car brochure collector, Thai food lover, not a morning person before my first cup of coffee

21 thoughts on “Shift Happens”

  1. Good morning, Bruno. Thanks for today’s article.

    I was aware of most of the transmission issues with the Traction Avant, but it was only after I read Gijsbert-Paul Berk’s book about André Lefebvre that I heard the name of Dimitri Sensaud de Lavaud. The anecdote about the fries is in there too, but it is not specifically linked to Lefebvre, but to someone who worked at the Bureau d’Etudes. Not that it matters of course.

    The Voisins got a mechanical gearbox complimented by an electromagnetic operated planetary gearbox from Cotal, after the Sensaud de Lavaud automatic gearbox proved to be too problematic.

    The gearbox might have been the Traction Avant Achilles’ heel, but what a great car that was and its success was much deserved.

  2. Good morning Bruno. I love reading about the people and personalities that shaped the automotive industry and this is a cracking tale! Many thanks for sharing.

  3. Traction Avant: A moral tale about the balance between Technic And Finances.

    Traction Avant.

    How you can bankrupt your company with a superb car. You can make a car TOO advanced and TOO sophisticated: Financial Gods are going to punish you.

    Ford England / Toyota in the 1970s

    How you can earn money with mediocre, unimaginative cars

    And I say this as the owner of a 2004 Avensis. I am a Toyota Converse, a preacher of the Toyota Gospel, BUT Toyota made technical junk in the 70s…profitable technical junk.

  4. Interesting story, of which I, as a reader of many Citroën-related books and articles, wasn’t aware yet. I wonder if anything about it is mentioned in my Citroën ‘Bible’ (“Toutes les Citroën”) and I simply read over it… Anyway, thanks for this piece!

    I only enjoyed a short (15 min) drive in a Traction once, and actually had my struggles with shifting. I don’t have any other experience with gearboxes of that time, but I found it quite hard to actually find the right spot to move the lever to and to engage gears without crashing noise. Compared to the smoothness and level of relaxation the car exuded in most other chapters (especially suspension and seating), it was indeed a mismatch.

    It seems that it’s a constant in Citroën’s flagship development that new components were developed unsuccessfully and had to be abandoned in the last minute. As with the boxer engine for the DS or Wankels for the CX. I wonder if there are other manufactureres with such a record of similar stories.

  5. “André Citroën finally had to face the facts, and instructed the Bureau d’Etudes to design and build a conventional manual gearbox for the Traction”

    The same thing happened four decades later with the Wankel engine and the GS/CX

    The (reliable) Wankel engines never arrived, so Citroen had yo improvise a conventional(ish) technical solución:

    And the engines were the Achilles heel of those Cars.

    The engine bay of those Cars was todo small… Because ir was designed for a Wankel engine that never came.

    And then came the SECOND bankruptcy of Citroen.

    1. The article behind the link to the wankel should be taken with care because it contains some technical as well as historical nonsense.

      There’s a big difference between the de Lavaud transmission and the wankel engine.
      The transmission served a useful purpose but was badly executed and could not be made to work, just like the boxer engine for the DS.
      The wankel was a fully developed technical blind alley and in the end it was an answer to a question nobody ever asked. That’s why VW stopped it and closed Comotor, putting an end to wankel projects based around the KKM871 engine.

  6. Explaining “automatic drive through a system of oscillating plateaus”. That sounds like a textbook example of An Ecumenical Matter.

    However, moving on to the later Sensaud de Lavaud transmission, all credit to the pragmatic Citroën engineers for preparing a ready-to-go alternative when the boss’s favoured gearbox turned out to be as problematic as expected. Such things often happen in the engineering and design world.

  7. Concerning the oscillating plateaus transmission: in my research for this article I did find a description of it on a French forum, which I translated to English by means of an online translation website. As even after re-reading it several times I still did not understand it I decided not to include the text in the article- but I will post it here; perhaps more mechanically inclined and versed DTW readers than me (Dave or Bob come to mind) can make sense of it?
    Here it is:

    The input shaft is driven by the motor. It drives a variable tilt plate (called a “swash plate”). This one can not turn because it is stopped by the slide, it just oscillates. This is the same principle as in a variable displacement hydraulic pump. The top carries six ball joints arranged in a star shape.
    On these ball joints are articulated connecting rods, the other of which end is hinged on the outer cages of six freewheels, the hubs of which are fixed on the axle (the scheme shows only one connecting rod, in reality the six connecting rods are articulated on one side on the plate, in a star shape; and on the other
    side on the lower part of the axle). When the tray is not tilted, the connecting rods are fixed and the car does not advance.
    When the tray tilts, the connecting rods begin to move. When they recoil, the freewheels lock and drive the axle. As they move forward, the outer cages unlock and spin on their respective hubs.
    The tilt of the platform is automatic, it depends on the speed of the engine and the resistant torque due to the profile of the road.
    Due to the design of the system (presence of free wheeling), there is no engine brake. An inverter is arranged between the freewheels and the differential (also freewheels, designed by Dimitri Sensaud de Lavaud).
    Sensaud de Lavaud transmissions were manufactured by the Société de Construction des Battignoles (SCB, now SPIE-Battignoles).
    In addition to Lavaud’s Sensaud car, this transmission was also tried at Voisin, where it did not leave good memories.
    However, the book also supplies a translation of an article by W. F. Bradley, published in “The autocar” two years earlier, who traveled 3200 km in the South-East of France without any problems.

    1. The whole thing looks like the combination of a helicopter swashplate and connecting rods working on the drive shaft.

      One of the innumerable alternative solutions to mechanical questions already answered. At least it uses power transfer by form instead of by friction like most other CVT designs.

    2. A complicated and expensive solution in search of a problem. A potential winner for the competitions for the most complicated mechanism to operate a light switch or to drop an egg…

      Do a thought experiment: imagine sitting on a bicycle with a freewheeling mechanism and (I’m sure you did it as a kid) try to move the bicycle by rocking the pedals back and forth for maybe 45 or 60 degrees instead of using them in a continuous circular motion. The bicycle will move but with an uneven spread of motive power. No imagine having not one, but six sets of pedals that are used in sequence, thereby generating a continuous flow of power through this ‘ratcheting’ motion.
      Now imagine conrods connected to the pedals and you move the ‘ratchets’ by pulling on the conrods. These conrods are arranged not in a row but in a circle and their free ends are attached to a disc rotating on a shaft – or rather two concentric discs of which the inner rotates with the shaft and the outer is fixed and does not rotate but carry the conrods just like the swashplate in a helicopter.
      As long as the disc surface is perpendicular to the shaft there will be no rocking motion on the conrods. Now tilt the disc for a couple of degrees and the disc will move the conrods for the distance between the point on the disc that is closest to the rods and the one that is most away from them. The more you tilt the disc the bigger the travel of the conrods and the bigger the angle these conrods operate the ratcheting mechanism driving the output shaft.

      Look at the video and imagine the rods being connected not to rotor blades but to the ratcheting mechanism.

    3. Thanks Dave for your clarification; I believe I have a better understanding of how this transmission worked (or was supposed to work). It does sound like a very complex method to achieve its goal, but taking into account that Mr Sensaud de Lavaud designed this almost a century ago and had to make do with what was known at the time in terms of materials and available mechanical devices I still respect his efforts both in this and the subsequent turbine transmission.

    4. The turbine transmission doesn’t sound like a too bad idea.
      It looks like a Foettinger fluid coupling or an early torque converter and the problem seems to have been that they tried to use ‘slip’ in the coupling to modulate the speed of the output shaft rather than use some gears to do this. If the TA’s gearbox uses the casing originally intended for the hydraulic transmission then there would not have been much room for epicyclic gears which you normally need because power transfer through the box cannot be interrupted thanks to the hydraulic coupling and therefore gears have to be changed under load. Either they used some gears of an unknown kind (no unknown unknowns, please) or none at all.
      Too much difference in speed between the elements of a fluid coupling of course heats up the oil.
      Couplings of this kind are used diesel railway locomotives but their engines operate at a much lower speed and the slip in the coupling is achieved by dynamically draining or filling the oil.
      Automatic transmissions normally have oil coolers to keep oil temperature in check.

  8. Might one infer that bankruptcy might have been averted, had Citroen not been seduced by the automatic transmission?

  9. Would Citroen have had a better chance of surviving with the Traction Avant had it featured a 3-speed if not a 4-speed gearbox from the beginning instead of the automatic, if so would it have also improved the prospects of the 3.8-litre V8 22 CV project reaching production (and indirectly finding its way into the DS in revised form)?

  10. A wonderful, enlightening article as usual, Bruno – thank you.

    I don’t think that I can help with the ‘oscillating plateaux’ transmission, but here’s a short video explaining the casting process. Dimitri clearly liked the application of centrifugal forces / turbine-like devices. I wonder what the Dynaflow transmission’s secret was which made it work – something like an oil cooler, perhaps?

    I’d love to learn more about the Traction Avant – I’ve heard it was advanced, but I’ve never managed to do enough research to understand why, properly.

    1. Thank you for your kind comment Charles; and for the video about the centrifugal casting process.
      i don’t know why GM/Buick managed to make in essence the same principle work satisfactory and especially reliably; the thought of fitting an oil cooler crossed my mind too- but the problem(s) of De Lavaud’s version must have been more deeply rooted than that as I can not imagine Citroën’s Bureau d’Etudes, let alone an excellent engineer like
      Lefebvre, not thinkinmg of that as well.

    2. I’m not technically minded, so I’m no help in explaining things (talk about a day late and a dollar short), but two possibilities spring to mind about the oil cooling: maybe the technology for that wasn’t advanced enough yet (I have no idea about timelines for this kind of technology), or maybe there simply wasn’t enough room in the casing to add such equipment?

      Other than that: wonderful story about a great car and great (if flawed) people.

  11. Thank you for this article, it is informative and the same time there is a sense of tension in the plot, like will the Citroen engineers design and accomplish a gearbox in time for this car, or not? There was small distance between success and failure. And I was thinking that in these times people lived and worked in slower rythms than today.

  12. Thank you the article, an enjoyable read. I found it searching for “Sensaud de Lavaud” having revisited my copy of “André Citroën: The Man and the Motor Cars” by John Reynolds, reading the chapter on the development of the Traction Avant. In fact my first car was a 12-year old British-built Light Fifteen, which I bought in 1965 while at university with the proceeds of my first paid summer job. Mine too suffered a catastrophic gearbox failure due to the loss of several crown wheel/pinion teeth. I replaced it with a whole second-hand transaxle unit which survived the next few years while I owned the car. Otherwise a great car, still ahead of many of the designs being sold in the fifties and sixties. There was very little that could touch it in winter conditions.
    Regarding the original Sensaud de Lavaud transmission, it sounds like the hydrostatic transmissions used in off-highway machinery of which Rexroth is a prime example. Here is an illustration:

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: